Integrating the technology infrastructures of two separate organizations after a merger or acquisition can be living hell for the IT departments involved. Ever tried to knit together two multiterabyte SANs -- one comprising Brocade Communications Systems Inc. (Nasdaq: BRCD) switches, and the other McData Corp. (Nasdaq: MCDTA) directors -- that are 20 miles apart, each supporting several thousand users, with only five people to complete the job?
Oil refining giant Valero Energy Corp. faced no choice when it merged with Ultramar Diamond Shamrock Corp. in December 2001. It had to find a way to consolidate its two disparate storage networks into one.
The company's two data center campuses, one in Valero's corporate office near downtown San Antonio and the other 20 miles north at the previous headquarters of Ultramar Diamond Shamrock, had two entirely different SANs. One consisted of five Brocade SilkWorm 2800 switches; the other included one McData 6064 and two ES-500 directors.
"And never the twain shall meet." That's what Kent Arnott, technical specialist at Valero, says was his first thought when he approached the problem.
Together, the SANs support 30 TBytes of data housed in Hitachi Data Systems (HDS) and IBM Corp. (NYSE: IBM) arrays connected to 300 servers. Valero needed something that would merge the two architectures, allowing centralized management of this disk space. In addition, in combining the two networks, Valero needed to provide support for multipath I/O host bus adapters (HBAs) for Sun Microsystems Inc. (Nasdaq: SUNW) Solaris, Windows NT and 2000, IBM AIX, and Hewlett-Packard Co. (NYSE: HPQ) UX; it needed the SAN to be fully redundant; and it needed to support HP Storage Data Protector clustering software and IBM Tivoli Storage Manager [ed. note: No walk in the park!].